There were multiple paths that could lead to a fast box
allocation. One of them made little sense (in-place send) so it has
been removed to allow a rework of the fast-box send function. This
should fix a number of issues with hanging/crashing when using the
vader btl.
References #4260
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The vader btl kept a per-peer registration cache to keep track of
attachments. This is not really a problem with small numbers of local
ranks but can be a problem with large SMP machines. To reduce the
footprint there is now one registration cache for all xpmem
attachments. This will probably increase the lookup time for large
transfers but is a worthwhile trade-off.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit fixes several bugs in the registration cache code:
- Fix a programming error in the grdma invalidation function that can
cause an infinite loop if more than 100 registrations are
associated with a munmapped region. This happens because the
mca_rcache_base_vma_find_all function returns the same 100
registrations on each call. This has been fixed by adding an
iterate function to the vma tree interface.
- Always obtain the vma lock when needed. This is required because
there may be other threads in the system even if
opal_using_threads() is false. Additionally, since it is safe to do
so (the vma lock is recursive) the vma interface has been made
thread safe.
- Avoid calling free() while holding a lock. This avoids race
conditions with locks held outside the Open MPI code.
Fixesopen-mpi/ompi#1654.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit rewrites both the mpool and rcache frameworks. Summary of
changes:
- Before this change a significant portion of the rcache
functionality lived in mpool components. This meant that it was
impossible to add a new memory pool to use with rdma networks
(ugni, openib, etc) without duplicating the functionality of an
existing mpool component. All the registration functionality has
been removed from the mpool and placed in the rcache framework.
- All registration cache mpools components (udreg, grdma, gpusm,
rgpusm) have been changed to rcache components. rcaches are
allocated and released in the same way mpool components were.
- It is now valid to pass NULL as the resources argument when
creating an rcache. At this time the gpusm and rgpusm components
support this. All other rcache components require non-NULL
resources.
- A new mpool component has been added: hugepage. This component
supports huge page allocations on linux.
- Memory pools are now allocated using "hints". Each mpool component
is queried with the hints and returns a priority. The current hints
supported are NULL (uses posix_memalign/malloc), page_size=x (huge
page mpool), and mpool=x.
- The sm mpool has been moved to common/sm. This reflects that the sm
mpool is specialized and not meant for any general
allocations. This mpool may be moved back into the mpool framework
if there is any objection.
- The opal_free_list_init arguments have been updated. The unused0
argument is not used to pass in the registration cache module. The
mpool registration flags are now rcache registration flags.
- All components have been updated to make use of the new framework
interfaces.
As this commit makes significant changes to both the mpool and rcache
frameworks both versions have been bumped to 3.0.0.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit fixes several threading bugs:
- Add an additional lock to the btl_base_endpoint_t structure to lock
the list of pending frags. This allows the progress function to
attempt to send pending frags without needing to drop/reaquire the
lock. This should provide a small improvement in performance and
fixes a potential race between adding an removing items from the
pending list.
- Ensure fast boxes are only set up once by updating the send count
using atomics when needed and do not set the fast box buffer
pointer until the fast box is set up.
Closesopen-mpi/ompi#1408
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Bring Slurm PMI-1 component online
Bring the s2 component online
Little cleanup - let the various PMIx modules set the process name during init, and then just raise it up to the ORTE level. Required as the different PMI environments all pass the jobid in different ways.
Bring the OMPI pubsub/pmi component online
Get comm_spawn working again
Ensure we always provide a cpuset, even if it is NULL
pmix/cray: adjust cray pmix component for pmix
Make changes so cray pmix can work within the integrated
ompi/pmix framework.
Bring singletons back online. Implement the comm_spawn operation using pmix - not tested yet
Cleanup comm_spawn - procs now starting, error in connect_accept
Complete integration
Use of the old ompi_free_list_t and ompi_free_list_item_t is
deprecated. These classes will be removed in a future commit.
This commit updates the entire code base to use opal_free_list_t and
opal_free_list_item_t.
Notes:
OMPI_FREE_LIST_*_MT -> opal_free_list_* (uses opal_using_threads ())
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The old BTL interface provided support for RDMA through the use of
the btl_prepare_src and btl_prepare_dst functions. These functions were
expected to prepare as much of the user buffer as possible for the RDMA
operation and return a descriptor. The descriptor contained segment
information on the prepared region. The btl user could then pass the
RDMA segment information to a remote peer. Once the peer received that
information it then packed it into a similar descriptor on the other
side that could then be passed into a single btl_put or btl_get
operation.
Changes:
- Removed the btl_prepare_dst function. This reflects the fact that
RDMA operations no longer depend on "prepared" descriptors.
- Removed the btl_seg_size member. There is no need to btl's to
subclass the mca_btl_base_segment_t class anymore.
...
Add more
using knem
It is valid to modify the remote segment that will be used with the
btl put/get operations as long as the resulting address range falls in
the originally prepared segment. Vader should have been calculating the
offset of the remote address in the registered region. This commit
fixes this issue.
We recognize that this means other users of OPAL will need to "wrap" the opal_process_name_t if they desire to abstract it in some fashion. This is regrettable, and we are looking at possible alternatives that might mitigate that requirement. Meantime, however, we have to put the needs of the OMPI community first, and are taking this step to restore hetero and SPARC support.
This commit makes the folowing changes:
- Add support for the knem single-copy mechanism. Initially vader will only
support the synchronous copy mode. Asynchronous copy support may be added
int the future.
- Improve Linux cross memory attach (CMA) when using restrictive ptrace
settings. This will allow Open MPI to use CMA without modifying the system
settings to support ptrace attach (see /etc/sysctl.d/10-ptrace.conf).
- Allow runtime selection of the single copy mechanism. The default behavior
is to use the best available. The priority list of single-copy mehanisms is
as follows: xpmem, cma, and knem.
- Allow disabling support for kernel-assisted single copy.
- Some tuning and bug fixes.
This is a large update that does the following:
- Only allocate fast boxes for a peer if a send count threshold
has been reached (default: 16). This will greatly reduce the memory
usage with large numbers of local peers.
- Improve performance by limiting the number of fast boxes that can
be allocated per peer (default: 32). This will reduce the amount
of time spent polling for fast box messages.
- Provide new MCA variables to configure the size, maximum count,
and send count thresholds for fast boxes allocations.
- Updated buffer design to increase the range of message sizes that
can be sent with a fast box.
- Add thread protection around fast box allocation (locks). When
spin locks are available this should be updated to use spin locks.
- Various fixes and cleanup.
This commit was SVN r32774.
WHAT: Merge the PMIx branch into the devel repo, creating a new
OPAL “lmix” framework to abstract PMI support for all RTEs.
Replace the ORTE daemon-level collectives with a new PMIx
server and update the ORTE grpcomm framework to support
server-to-server collectives
WHY: We’ve had problems dealing with variations in PMI implementations,
and need to extend the existing PMI definitions to meet exascale
requirements.
WHEN: Mon, Aug 25
WHERE: https://github.com/rhc54/ompi-svn-mirror.git
Several community members have been working on a refactoring of the current PMI support within OMPI. Although the APIs are common, Slurm and Cray implement a different range of capabilities, and package them differently. For example, Cray provides an integrated PMI-1/2 library, while Slurm separates the two and requires the user to specify the one to be used at runtime. In addition, several bugs in the Slurm implementations have caused problems requiring extra coding.
All this has led to a slew of #if’s in the PMI code and bugs when the corner-case logic for one implementation accidentally traps the other. Extending this support to other implementations would have increased this complexity to an unacceptable level.
Accordingly, we have:
* created a new OPAL “pmix” framework to abstract the PMI support, with separate components for Cray, Slurm PMI-1, and Slurm PMI-2 implementations.
* Replaced the current ORTE grpcomm daemon-based collective operation with an integrated PMIx server, and updated the grpcomm APIs to provide more flexible, multi-algorithm support for collective operations. At this time, only the xcast and allgather operations are supported.
* Replaced the current global collective id with a signature based on the names of the participating procs. The allows an unlimited number of collectives to be executed by any group of processes, subject to the requirement that only one collective can be active at a time for a unique combination of procs. Note that a proc can be involved in any number of simultaneous collectives - it is the specific combination of procs that is subject to the constraint
* removed the prior OMPI/OPAL modex code
* added new macros for executing modex send/recv to simplify use of the new APIs. The send macros allow the caller to specify whether or not the BTL supports async modex operations - if so, then the non-blocking “fence” operation is used, if the active PMIx component supports it. Otherwise, the default is a full blocking modex exchange as we currently perform.
* retained the current flag that directs us to use a blocking fence operation, but only to retrieve data upon demand
This commit was SVN r32570.
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
